User Requirement Analysis for a Real-Time NLP-Based Open Information Retrieval Meeting Assistant


Alcaraz B., Hosseini-Kivanani N., Najjar A., Bongard-Blanchy K.


Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13980 LNCS, pp. 18-32, 2023


Meetings are recurrent organizational tasks intended to drive progress in an interdisciplinary and collaborative manner. They are, however, prone to inefficiency due to factors such as differing knowledge among participants. The research goal of this paper is to design a recommendation-based meeting assistant that can improve the efficiency of meetings by helping to contextualize the information being discussed and reduce distractions for listeners. Following a Wizard-of-Oz setup, we gathered user feedback by thematically analyzing focus group discussions and identifying this kind of system’s key challenges and requirements. The findings point to shortcomings in contextualization and raise concerns about distracting listeners from the main content. Based on the findings, we have developed a set of design recommendations that address context, interactivity and personalization issues. These recommendations could be useful for developing a meeting assistant that is tailored to the needs of meeting participants, thereby helping to optimize the meeting experience.



Share this page: